diff --git a/README.md b/README.md
index 953ad15..692b62c 100644
--- a/README.md
+++ b/README.md
@@ -1735,6 +1735,10 @@ multi-host training can make it very easy to introduce bugs!*
     trials to run in parallel and we can afford to run many trials in sequence,
     Bayesian optimization becomes much more attractive despite making our tuning
     results harder to interpret.
+-   Both quasi-random search and Bayesian optimization can be used together with
+    [Hyperband](https://jmlr.org/papers/volume18/16-558/16-558.pdf) /
+    [ASHA](https://arxiv.org/abs/1810.05934), which quickly discard points that
+    aren't promising from the search.
 
 [^3]: Ben Recht and Kevin Jamieson
     [pointed out](http://www.argmin.net/2016/06/20/hypertuning/) how strong